Education System Intervention Modeling Framework
1 . Introduction and Background
Every year , millions of dollars are
spent on education interventions , and every year , many of those initiatives prove to be unsuccessful , unsustainable , or not scalable . From federal policies such as No Child Left Behind ( 2001 ), to research grants , to efforts of individual teachers and administrators , it is difficult to measure the outcomes of educational interventions and often more difficult to understand why they fail or are not sustainable ( Fullan , 2000 ). In spite of the complexity of school systems , most educational policymakers and reformers do not rely on models to make decisions about where to intervene and how to manage resources during interventions . Such decisions often have unintended consequences or outcomes ( Lubienski , 2005 ). Models of schools themselves , when designed carefully and used appropriately , can provide valuable insights about factors affecting intervention success and sustainability .
Education researchers , public policy experts , and complexity experts have studied different aspects of education systems ; however , these three disciplines have rarely collaborated in order to get a more complete picture . Education research as a field has broadened from simply designing and evaluating interventions to studying implementation within the school system ( Fixsen , Naoom , Blase , Friedman , & Wallace , 2005 ). This work has identified several factors that affect intervention implementation and outcomes , including professional development ( PD ), leadership , organization , school structure , resources , and support ( Billig , Sherry , & Havelock , 2005 ; Blumenfeld , Fishman , Krajcik , Marx , & Soloway , 2000 ; Ely , 1990 ; Elmore , 1996 ; McLaughlin & Talbert , 2003 ; Spillane , Reiser , & Reimer ,
2002 ). Social network analysis ( SNA ) has been applied to understand how a school ’ s social structure and teacher networks affect intervention implementation ( Coburn , Russell , Kaufman , & Stein , 2012 ; Daly , 2010 ; Moolenaar , 2012 ; Moolenaar & Daly , 2012 ). Moreover , experts in educational and public policy have studied reform agendas within social and organizational contexts ( Borman , Carter , Aladjem , & LeFloch , 2004 ; Crawford & Ostrom , 1995 ; Weaver- Hightower , 2008 ). While these approaches aid in understanding education interventions and policies in general , they do not capture the underlying mechanisms and feedback loops affecting intervention implementation and sustainability .
Meanwhile , engineers and complexity experts have only recently turned their attention to the education realm . Recent efforts have applied system dynamics ( SD ) and agent-based modeling ( ABM ) to understand student interest and selection in STEM ( Science , Technology , Engineering and Mathematics ) ( Allen & Davis , 2010 ; Sanchez , Wells , & Attridge , 2009 ). ABM has emerged as a popular technique for modeling social systems because it captures emergent behaviors , is a natural description of a system , and is flexible enough to accommodate different scales temporally and spatially ( Bonabeau , 2002 ). Other researchers have promoted the use of SD for studying education as a complex system ( Groff , 2013 ). These models rely on survey data to formulate causal relationships , but often lack a mechanism for distinguishing correlation from causation ( Pedamallu , Ozdamar , Akar , Weber , & Özsoy , 2012 ; Pedamallu , Ozdamar , Ganesh , Weber , & Kropat , 2010 ). To understand the effects of an intervention on a particular school system , systems engineering , policy , and education research approaches need to be combined , leveraging SD , ABM , and SNA
74